Beruflich Dokumente
Kultur Dokumente
ABSTRACT
Cloud computing is rising technology which is a new ordinary of large scale distributed computing. Cloud
Computing associates to both the applications delivered as a service and datacenters, where hardware&
software existing to provide service to customers through internet. A Cloud Computing is growing rapidly and
more users are attracted towards a fast service needs to provided. The Load can be memory network or delay
load. It always make necessary share work load among various nodes of the distributed system to improve the
resource utilization, maximum throughput, and maximum response time and for better performance of the system.
Here in this paper we have study many different Load Balancing Algorithms in Cloud Computing.
Keywords:- Cloud Computing; Load Balancing; Load Balancing Algorithms.
I.
INTRODUCTION Data as a service (DaaS) is also called as a cloud.
Cloud Computing is achieving a computing
tasks via network connection at a same time
1.1. Cloud Computing remaining in inaccessible from the complex
computing hardware and software infrastructure.
The term ‘cloud’ has ancient been used in the
telecommunication industry as a concept of the network in system [Yoganandani et
a representation for authorize all over Fig. 1. Cloud Computing [ Patidar et al.,(2012) ].
Convenient on-demand network access to
shared pool arrangement of computing
resources(example network server, storage
applications and services) that can be
immediately need and released with essential
micromanagement effort or service provider
communication.
Cloud
Cluster Grid Utility SaaS PaaS IaaS
Computing
Computing Computing Computing
I
The various essential characteristcs of cloud computing is On-Demand Self Service,Broad Network Access
,Resource Pooling ,Rapid Elasticity, and Measured Service are shows in Fig.3.
On-Demand Self-Service A user can utilize computing services Email, Applications resorting to human
interaction with [Yoganandani et al.,(2014)] providers of these resources.
Broad Network Access The computing resources are available over the network and accessed through
standard mechanisms that promote use by heterogeneous thick or thin client platforms such as mobile
phones, laptops situated at customer location.
Resource Pooling A cloud computing resources are shared together to serve multiple customers using either
multi tenancy.
Rapid Elasticity Cloud computing separately from traditional datacenter .Multiple Occupants occur in a cloud
environment that share components of a shared resource pool.
Measured Service Metered services also referred to as “pay per use” in a mean of responsibility. Resources
utilization is measured by check storage usage, CPU hours and bandwidth usage etc.
On –Demand Service
Resource Pooling
Rapid
Department of
Elasticity
Fig . 3. Characteristics Cloud Computing [MuthLakshmi,( 2013)].
Public Cloud Model The Public Cloud Model allows systems and services to be easily accessible to
public. Google, Amazon, Microsoft offers cloud services through internet are public cloud models. It may be
owned, managed, and operated by a business, academic, or government organization or some combination of them
also has several advantages which make the people to use easily. But few limitations also exist which are low
security and Less customizable.
Private Cloud In this private cloud model, it allows the systems and services to be accessible only by the
people in the organization which means other than the people in the specified organization are not allowed to use
access the services.
Community Cloud It may be managed, operated and owned by one or more of the organizations in the
community.
Hybrid Model Is an environment which combines public clouds and private clouds by allowing data and
applications to be shared between them.
SaaS (Software-as-a-Service) In this model a software provider license a software application to be used
and purchase on demand in multi-tenant environment, runs on web browser in client machine. Example Gmail
and Salesforces.com.
PaaS (Platform-as-a-Service) It offers a development platform for both completed and in progress cloud
application offer an environment where developer can develop to create and deploy applications and do not need to
know how much memory and processors their application requires. Example Google App Engine, Java,
Ruby Language.
IaaS (Infrastructure-as-a-service) IaaS service provider offer virtual server containing one or more CPU
running several choices of operating (IaaS). Example Web Services offer IaaS.
Software as a Platform as a Infrastructure as a
Service (SaaS) Service (PaaS) Service (IaaS)
highest priority [Ravi et al.,(2015)].This separate queue for each node. In load balancing
algorithm the process begins with considerate the with increasing and decreasing web server’s
information about the execution time of all demands. This mechanisms in virtual server
resources and then among all the resources a and load balancing is also useful while occupy
node with larger execution time will be selected. the server for a process. Advantages are
maximizing the throughput and overhead for
3.1.5. Throttled Load balancing Algorithm In system become minimum.
this algorithm end user first request the load 3.1.8. Genetic Algorithm This algorithm is used
balancer to check the availability of VM when soft computing approach [Rajeshkannan and
the VM is set to be free the next job is assigned Aramudhan,(2016)] .This algorithm provides a
to it and is dynamic in nature [Ravi et better performance, compare to RR and
al.,(2015)]. It helps in achieving better FCFS algorithm. GA’S implementation
performance and high utilization of resources. mechanism is based on three steps.
Initialization Operator Initial population is
3.1.6. Ant Colony Optimization Algorithm Ant
that the set of all population that are
Colony Optimization algorithm simulates the ant
employed in the genetic algorithm.
foraging behavior. In this algorithm the behavior
Selection Operator Best chromosome selects from
of ants is used for gathering information from
the parents and produce offspring.
different nodes in the system. When execution
Crossover Operator It combines two chromosomes
begin the ant and its [Ravi et al.,(2015)] that produce new chromosomes produce new offspring.
pheromone will get proposed from the head node
Mutation Operator It Changes of one or
and moves to the next node .If the ant finds any
more then gene value in a chromosomes. Or
of the nodes under loaded it will move forward to
when blunt change in a given population
another node and if that node is overloaded it will
chromosomes.
come back to previous node .
3.1.9. Particle Swarm Optimization Algorithm
Working of ACO Algorithm PSO is a swarm based heuristic optimization
technique [Bansal and Kaur,(2016)]. It is used
Ant Generation Analyze the cloud platform
for analyzing optimal path of solution space
[Verma et al.,(2017)] if there are overloaded or
while putting up load on a specific VM for
under loaded nodes, only then ants are generated
processing of resources it moves along all the
trying to minimize the makespan and maintain
VM and determinate the optimal machine to put
load across all the nodes with the given better
the load. It is one of the mechanisms to identify
results.
the optimal VM which is load less, and
To find target node The ant is looking for
available .So the relative energy and time
the target nodes with suitable condition of load
utilization to the node can be reduced.
balancing in its surrounding area.
3.1.7. Honeybee Foraging Algorithm This
algorithm is inspired from the behavior of
honeybees that uses the method to find and reap
food from the sources. After they come back to
honey comb and computes the food left, if there
is enough amount of food they will halt in the
comb in another way they will go out in search of
more honey [Ravi et al.,(2015)][ Nema,(2016)]
. In this algorithm resources are grouped as
virtual resources. Each virtual resource
maintains a process queue and accepts each
request from the queue and processes those
requests. That is why this algorithm is named as
Honey bee foraging algorithm. Maintaining a
Table. 1. Comparison between various existing Load Balancing in Cloud Computing [Malani and Amdani,(2015)].
minimum task completion and task response Subasish Mohapatra et al. (2013) This
time. The proposed algorithm overcome shows different load balancing algorithms in
disadvantages of honey bee inspired algorithm is order to manage the resources of services
it select tasks in first come first order not provider efficiently and effectively. This paper
analysis of all the free VM [Baby,(2014)] . presents a comparison of various policies
Shanti Swaroop Moharana et al. (2013) utilized for load balancing. Future work based on
Author showed that the cloud computing is a this modified algorithm better response time can
new technology which uses VM replacement of be expected if we apply some evolutionary
Physical machine to host the different algorithms such as PSO, ACO instead of
components. Load balancer are used for different scheduling algorithms [Mohapatra et al.
assigning load to different VM along none of ,(2013)].
the nodes get loaded heavily or lightly Tran Cong Hung, Nguyen Xuan Phi (2016)
overloaded of system lead to poor performance Author presents that more effective methods of
but our focus is on various load balancing load balancing, in order to increase system
algorithms and applications performance. Therefore in thisin cloud
papercomputing
we environment [ Mo
Hussain A Makasarwala, Prasun Hazari researched some parameters affecting the
(2016) The author shows the load balancing is performance load balancing on the cloud
one of the major issues in cloud computing. But computing and reduce the time makespan of VM [
this paper gives a Genetic Algorithm approach Hung and Phi,(2016)]. Navtej Singh
for load balancing in cloud. The idea behind in Ghumman, Rajesh Sachdeva (2016) Author
considered the priority is to get real world describes an effective and efficient and
virtualization. Simulation of the proposed method enhanced composite scheduling algorithm that
is done using Cloud Analyst. It will give better can be used to maintain the load and provides
response time compared to previous available efficient resource allocation techniques. In these
methods [Makasarwala and Hazari,(2016) ]. paper composite approaches is applied for load
Y. Kaushik et al. (2015) Author presents balancing using Throttled Algorithm and
an enhanced algorithm through compression Equally Spread Current Execution (ESCE)
various cloud load balancing algorithms. This Algorithm and take less response time and
paper shows the comparison of various load average time to achieve the objective [Ghumman
balancing algorithms like Round Robin, Equally and Sachdeva,(2016)].
Spread Current Execution load balancing and Kousik Dasgupta et.al (2013) Author
Throttled in cloud with different metric like describes a new load balancing strategy
Response time. Throttled Algorithm allocates using Genetic Algorithm (GA).This algorithm to
the resources to the job in queue leading balance the load of the cloud infrastructure while
reduced cost in data transfer and VM machine trying minimizing the completion time of a given
formation [Kaushik et al.(2015)]. task set. Load balancing action has been
Maysoon A.Alamin et al. (2017) In this simulated using the Cloud Analyst Simulator.
author presents the challenging problems that GA has been used however variation of the
affect the load balancing process is scheduling crossover and selection strategy could be applied
incoming request in an efficient way with as a future work for getting more better results.
minimum response time .The new load PBO is better than GA if system response time is
balancing proposed algorithm handle the load not considered not only reduce the response time
balancing problem in cloud computing .This but also minimize makespan and enhance system
proposed algorithm combine the Throttled and performance[Dasgupta et al.,(2013)] .
Equally Spread Current Execution algorithm
results showed the proposed algorithms V. CONCLUSION
improve the cloud system performance by
decreasing the response time and cost Load Balancing is one of the main issue in
[Alamin(2017)]. cloud computing. Main purpose of load
balancing is to distribute load among various [5] Mishra Kumar. N.; Mishra Nishchol.
individual servers. In this paper, various (2015): Load Balancing Techniques
techniques related to efficient Load Balancing Need Objectives and Major
are analyzed and compared to the other Challenging in Cloud Computing -A
techniques. And we have also discussed Systematic Review. International Journal
required qualitative metrics for load balancing. In of Computer Applications, 131(18),
future work the soft computing hybrid pp.11-19.
techniques may increase the performance of the
system and to meet the requirements of load [6] Ravi . V; Kanakala Teja; Reddy Krishna
balancing in cloud computing. .V. (2015): Performance Analysis of
Load Balancing Techniques in Cloud
Computing Environment. Telkomnika
VI. FUTURE SCOPE
Indonesian Journal of Electrical
In future work, we are planning to optimize Engineering,13( 3), pp.568-573.
PSO algorithm to make it appropriate for cloud
[7] Nema Lekha; Sharma Avinash; Jain
computing and more efficient in terms of load
Saurabh.(2016):Load Balancing
balancing. In addition to this, the optimization
Algorithms in Cloud Computing.
of PSO on various cloud simulators and
International journal of Engineering
compare the proposed approach with
Science and Computing, 6(6),pp. 7463-
previously soft computing techniques based on
7468.
fixed parameters. It may increase the
performance of the system and to meet the [8] Verma .P; Shrivastava Sonika;
requirements of load balancing in cloud Patriya.K. R.2017): Enhancing Load
Balancing in Cloud Computing by Ant
computing. Colony Optimization Method.
International Journal of Computer
REFERENCES Engineering in Research Trends, 4(6),
pp. 277-284.
[1] oganandani ,et al. (2014): Clearing The [9] Rajeshkannan.R;Aramudhan.M.(2016):C
Clouds On Cloud Computing. omparative Study of Load
Balancing Algorithms in Cloud
International Journal of Recent Computing Environment.Indian Journal
Development in Engineering and of Science and Technology, 9(20),pp.2-7.
Technology. 1(1), pp. 117-118.
[10] Kaur Amanpreet ; Bansal.G.(2016):
[2] MuthLakshmi .V.N. (2013) .Cloud Enhancement of Dynamic Load
Computing-An Overview. Publication of Balancing Using Particle Swarm
Problems &Application in Engineering Algorithm in Cloud Environment.
Research, 4(1), pp.324-325. Journal of Computer Engineering, 18 (
4), pp.06-10.
[3] Pathak. R. (2017): A Survey of Load
Balancing in Cloud Computing [11] Amdani. S.Y; Malani .S.S.(2015) : A
Challenges and Algorithms. Study of Load Balancing in Cloud
International Journal of Innovative Computing using cloud partitioning
Research in Computer and and ACO. International Journal of
Communication Engineering, 5(1), Engineering and Technical Research,
pp.740-742. 3(4), pp.333-335.